The authors propose a mechanism that combines preferences and predictions for public project decision-making, ensuring robustness to manipulation.
